人工智能辅助宫颈液基细胞学诊断可行性的多中心研究  

Feasibility multi-center study of artificial intelligence assistance in cervical fluid-based cytology diagnosis

作　　者：吕京澴[1] 樊祥山[2] 沈勤 王晓骁[4] 章宜芬[5] 黄文斌[6] 曹益陆 周超 常江龙 马威 周晓军 张丽华[9] Lyu Jinghuan;Fan Xiangshan;Shen Qin;Wang Xiaoxiao;Zhang Yifen;Huang Wenbin;Cao Yilu;Zhou Chao;Chang Jianglong;Ma Wei;Zhou Xiaojun;Zhang Lihua(Department of pathology,Suzhou Municipal Hospital,Suzhou 215002,China;Department of Pathology,the Affiliated Drum Tower Hospital of Nanjing University Medical School,Nanjing 210008,China;Department of Pathology,Jinling Hospital,Nanjing 210002,China;Department of GCP Center,Jiangsu Province Hospital ofTCM,Nanjing 210029,China;Department of Pathology,Jiangsu Province Hospital ofTCM,Nanjing 210029,China;Department of Pathology,the Affiliated Nanjing Hospital of Nanjing Medical University,Nanjing 210000,China;Jiangsu Yitou Health Technology Company,Nanjing 210000,China;891360 Medical Technology Company,Nanjing 210000,China;Department of pathology,Affiliated Zhongda Hospital,Southeast University,Nanjing 210009,China)

机构地区：[1]南京医科大学附属苏州医院苏州市立医院病理科,苏州215002 [2]南京大学医学院附属鼓楼医院病理科,210008 [3]解放军东部战区总医院病理科,南京210002 [4]江苏省中医院临床试验中心,南京210029 [5]江苏省中医院病理科,南京210029 [6]南京医科大学附属南京医院病理科,210000 [7]江苏易透健康科技有限公司,南京210000 [8]玖壹卷陆零医学科技南京有限公司,210000 [9]东南大学附属中大医院病理科,南京210009

出　　处：《中华病理学杂志》2021年第4期353-357,共5页Chinese Journal of Pathology

基　　金：南京市卫生科技发展课题(YKK19066);南通市民生科技重点项目(MS22018013)。

摘　　要：目的本文采用基于深度卷积神经网络的方法,针对人工智能在宫颈液基细胞片病理图像自动筛查中的应用价值,开展多中心的实际应用研究,并与细胞学医师的诊断进行比较及分析。方法采用深度分割网络提取5 516张细胞学病理图像中的感兴趣区域618 333个,结合医师的经验训练出具有分析能力的深度分类网络,利用其分类结果构建特征,使用决策模型完成细胞病理图像的分级。结果该方法对4 908例宫颈液基细胞片进行病理图像自动筛查,灵敏度为89.72%,特异度为58.48%,阳性预测值为33.95%,阴性预测率为95.94%。在4种不同制片或染色方法的细胞片中,本算法对于巴氏染色自然沉降片效果最佳,灵敏度为91.10%,特异度为69.32%,阳性预测值为41.41%,阴性预测值为97.03%。结论深度卷积神经网络图像识别技术可初步应用于宫颈细胞学筛查。Objective To propose a method of cervical cytology screening based on deep convolutional neural network and compare it with the diagnosis of cytologists.Method The deep segmentation network was used to extract 618333 regions of interest(ROI)from 5,516 cytological pathological images.Combined with the experience of physicians,the deep classification network with the ability to analyze ROI was trained.The classification results were used to construct features,and the decision model was used to complete the classification of cytopathological images.Results The sensitivity and specificity were 89.72%,58.48%,33.95%and 95.94%respectively.Among the smears derived from four different preparation methods,this algorithm had the best effect on natural fallout with a sensitivity of 91.10%,specificity of 69.32%,positive predictive rate of 41.41%,and negative predictive rate of 97.03%.Conclusion Deep convolutional neural network image recognition technology can be applied to cervical cytology screening.

关 键 词：人工智能 细胞学技术 卷积神经网络 

分 类 号：R737.33[医药卫生—肿瘤]